JOURNAL OF VIROLOGY, Sept. 1991, p. 4839-48460022-538X/91/094839-08$02.00/0Copyright X 1991, American Society for Microbiology

Endogenous Interferon Specifically Regulates Newcastle DiseaseVirus-Induced Cytokine Gene Expression in Mouse Macrophages

RAINER ZAWATZKY,1* HERIBERT WURMBAECK,1 WERNER FALK,2 AND ANGELA HOMFELD'tInstitut fur Virusforschung1 and Institut fur Immunologie und Genetik,2

Deutsches Krebsforschungszentrum, D-6900 Heidelberg, Germany

Received 22 February 1991/Accepted 18 June 1991

In macrophages from inbred mice, the magnitude of the interferon (IFN) response to Newcastle disease virus(NDV) infection is under genetic control of the If-i locus, which carries the allele for either high (h) or low (1)IFN production. Here, we report that the activity of genes within the If-i locus is influenced by macrophage-derived endogenous IFN. In addition to various other biological effects, we observed that endogenous IFNspecifically downregulated NDV-induced IFN and interleukin 6 production. Preculture of bone marrow-

derived macrophages (BMM) from BALB/c (If-i') mice in macrophage colony-stimulating factor plusanti-WFN-j provoked a 30- to 50-fold increase in NDV-induced cytokine production compared with inducedcontrol cultures in macrophage colony-stimulating factor alone, whereas only a 4- to 6-fold increase wasobserved in anti-IFN-j-treated BMM from C57BL/6 (If-i) mice. This resulted in nearly complete abrogationof the genetically determined difference in the response to NDV. The increase was specific for NDV and wasmarked by strong additional activation of IFN-a genes. Studies using BMM from B6.C-H28c If-i congenicmice gave results identical to those obtained with BALB/c BMM. Addition of 20 IU of recombinant IFN-a4 toanti IFN-,-treated macrophages from B6.C-H28c mice 20 h prior to NDV infection strongly downregulated theIFN-ax, IFN-1, and interleukin 6 responses. The genetic difference between macrophages from If ih and If-i'mice was thus reestablished, since the same treatment caused only weak reduction of NDV-induced cytokinegene expression in BMM from C57BL/6 mice. These data suggest that the If-I* and If-i' alleles harborIFN-inducible genes that, foliowing activation, specifically suppress subsequent cytokine gene expression inresponse to NDV.

Virus-induced interferon (IFN) production in a host isunder genetic control. Studies using inbred strains of micehave provided evidence for the existence of several loci thatspecifically influence the amount of alpha/beta interferon(IFN-a/W) produced following infection with one particularvirus (for reviews, see references 12 and 13). These loci,named Ifloci, have high- and low-producer alleles which areinducer specific in that they influence IFN production onlyafter contact with the corresponding inducer. The various Ifloci are characterized by their quantitative effects on IFNsynthesis that lead to differences in IFN production between3- and 30-fold, but their mode of action and gene productsare not known. The If-] locus, which specifically influencesIFN production by Newcastle disease virus (NDV), harborstwo alleles, If-P' and If-i', coding for high and low IFNproduction, respectively. C57BL/6 mice harbor the high-responder allele and BALB/c mice harbor the low-responderallele, resulting in 10- to 20-fold differences in NDV-inducedcirculating IFN levels. In vitro studies aimed at character-izing the cell populations involved in IFN production re-vealed that cultures of resident peritoneal macrophages fromC57BL/6 and BALB/c mice also expressed the correspond-ing If-i allele when induced by NDV (15). Analysis of IFNproduction in a variety of congenic lines generated byrepeated backcrossing of F1 progeny with C57BL/6 parentalmice led to the identification of the particular line B6.C-H28c, carrying the If-i' allele and histocompatibility locusH-28 from BALB/c on a C57BL/6 genetic background, since

* Corresponding author.t Present address: Heinrich Pette-Institut, D-2000 Hamburg, Ger-

many.

NDV-induced IFN levels in these mice were similar to thoseobserved in BALB/c mice (14). In the present study, we usedcultures of bone marrow-derived macrophages (BMM) fromthese mouse strains to study in more detail the regulatorymechanisms of the If-1 locus on NDV-induced IFN produc-tion. Our results indicate that the If-i locus also influencesinterleukin 6 (IL-6) production in response to NDV and thatthe If-I and If-i' alleles are differentially influenced bymacrophage-derived endogenous IFN.

MATERIALS AND METHODS

Mice. Male C57BL/6 and BALB/c mice were purchasedfrom the Zentralinstitut fur Versuchstierkunde, Hannover,Germany. Male B6.C-H28c/By and additional C57BL/6 andBALB/c mice were obtained from the Jackson Laboratories,Bar Harbor, Maine. Throughout the experiments, 10- to14-week-old animals were used.Macrophage cultures. BMM were obtained as described

previously (23). The cells were cultured in RPMI 1640-10%fetal calf serum (FCS) supplemented with 500 CFU of eithercrude or purified mouse macrophage colony-stimulating fac-tor (M-CSF) per ml prepared from L-cell conditioned me-dium (LCM; kindly provided by E. R. Stanley; 42). Whenindicated, 100 neutralizing units (NU) of either monoclonalanti-mouse alpha interferon (anti-IFN-a) or monoclonal anti-mouse beta interferon (anti-IFN-P) (kindly provided by Y.Watanabe; 26) was added to the culture medium. Experi-ments were routinely carried out on day 7 or 8, whencultures consisted of more than 95% macrophages as ana-lyzed by F4/80 antigen expression (22). LCM was continu-ously monitored for the presence of antiviral activity. Afterfivefold concentration by Amicon ultrafiltration, LCM was

4839

Vol. 65, No. 9

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

4840 ZAWATZKY ET AL.

screened for induction of an antiviral state and of the 2'-5'oligoadenylate synthetase (OASE) in mouse embryo fibro-blasts. Controls included LCM to which 250 NU of mono-clonal anti-IFN-a, anti-IFN-0, or both had been added 1 daybefore. In no case, however, did we obtain evidence for thepresence of IFN activity in LCM.

Viruses. NDV strain Ulster (38) was isolated from infectedembryonated eggs as previously described (23). Virus titerswere determined in a standard hemagglutination assay usingchicken erythrocytes and are given as hemagglutinationunits (HAU). The NDV preparation used in this study had atiter of 3,600 HAU.Herpes simplex virus type 1 (HSV) strain WAL was

prepared from infected RC-37 cells as previously described(48).

Purification of monoclonal anti-IFN antibodies. Monoclonalantibodies to mouse IFN-a and IFN-P were prepared fromculture supernatants of hybridoma lines 4EA1 and 7FD3,respectively. The cell lines were a generous gift of Y.Watanabe (26). Dense, confluent hybridoma cultures wereobtained after growth in RPMI 1640-10% FCS and incubatedfor an additional 3 days in fresh RPMI 1640 without serum.The medium was then recovered and concentrated 100-foldfollowing ammonium sulfate precipitation. The antibodieswere then purified by high-performance liquid chromatogra-phy (Beckman System Gold) using rabbit anti-rat immuno-globulin G coupled to Diasorb AC 1500 (Diagen). Themonoclonal antibodies eluted at pH 3. The neutralizationtiters of the purified preparations from 4EA1 and 7FD3supernatants were 1:8,000 against 4 IU of recombinantIFN-a4 (rIFN-a4) and 1:24,000 against 4 IU of natural mouseIFN-P, respectively.Mouse rIFN-a4 was kindly provided by E. C. Zwarthoff

(46), and a preparation of natural mouse IFN-P (lot no.82011; specific activity, 1.3 x 107 IU/mg) was purchasedfrom Stratech Scientific Ltd.

Induction and determination of IFN and IL-6 activities incultures of BMM. On day 7 of culture, BMM were detachedfrom petri dishes with a rubber policeman following incuba-tion with ice-cold saline, reseeded into 24-well plates at adensity of 2.5 x 105 per well, and allowed to attach for 4 to6 h. They were then incubated with NDV or HSV prepara-tions for 1 h and refed, after removal of the inducer, withfresh, prewarmed RPMI 1640-5% FCS. After overnightincubation, supernatants were removed for IFN and IL-6assays. For these experiments, purified mouse M-CSF (kind-ly provided by E. R. Stanley; 42) was used in general, butidentical results were obtained with crude LCM as a sourceof M-CSF. IFN and IFN neutralization tests to determinethe relative proportions of IFN-a and IFN-3 were performedas previously described (23, 49). IL-6 activity was deter-mined in a colorimetric proliferation assay (35) using sub-clone B9.9 isolated from hybridoma cell line B13.29, kindlyprovided by L. A. Aarden (1). One unit of IL-6 was definedas giving half-maximal proliferation of B9.9 cells. Superna-tants to be assayed for IL-6 were preincubated with excessantibodies to mouse IFN-a/3 to neutralize the antiprolifera-tive effect of mouse IFN on B9.9 cells.DNA probes. Genomic DNA probes for mouse IFN-a1,

IFN-a2, and IFN-a4 and cDNA probes for IFN-P and IL-6were kindly provided by E. Zwarthoff (50), Y. Kawade (21),and J. van Snick (47), respectively. A 2.5-kb probe codingfor the F protein and part of the HN protein of NDV waskindly provided by P. T. Emmerson (8, 9). For RNA blotanalysis, the probes were labeled with [a-32P]dCTP byrandom priming (17).

Preparation of RNA and blot hybridization. Total cellularRNA was extracted from mouse BMM with 4 M guanidiniumthiocyanate as described by Chirgwin et al. (10). DenaturedRNA was electrophoresed through agarose gels in the pres-ence of formaldehyde (31) and transferred to nylon filters.Hybridization was carried out as previously described (23).For monitoring of identical amounts of cellular RNA per sloton the filter membrane, probes were removed by shaking thefilters in a water bath at 70°C in 1.5 mM NaCl-0.01% sodiumdodecyl sulfate, followed by rehybridization to labeled p-ac-tin cDNA (23).

Assay of OASE. IFN-induced, double-stranded RNA-de-pendent OASE activity was determined as described previ-ously (23). Briefly, the cells were lysed in a buffer containing0.5% Nonidet P-40 and protein extracts were bound topoly(IC) agarose (Pharmacia) and incubated with buffercontaining ATP (43). The 2'-5' oligo(A) produced was quan-titated by a competition radioactivity binding assay (27).

Vesicular stomatitis virus (VSV) infection and virus titra-tion. Macrophages were detached from petri dishes on day 7of culture and plated on 35-mm-diameter petri dishes at aconcentration of 5 x 105 per dish. At 6 h later, the cells werewashed and infected with VSV (strain Indiana) in serum-freemedium by using a multiplicity of 2 PFU per cell. Afteradsorption for 1 h at 37°C, the cell monolayer was washedagain to remove nonadsorbed virus particles and fresh,prewarmed RPMI 1640 plus 5% FCS was added. To accountfor residual input virus, an aliquot of the culture fluid wasremoved 1 h later. At 14 h later, culture dishes were frozenat -70°C to allow for approximately one replicative cycle ofVSV. Virus titrations were done by using a standard viralplaque assay on a monkey kidney cell line as previouslydescribed (29). The amount of input virus varied between 0.5x 103 and 1.5 x 103 PFU per culture and was deduced fromthe titers in the experimental groups.

RESULTS

Detection of endogenous IFN in cultures of BMM in thepresence of M-CSF. In cultures of BMM, endogenous IFN issecreted constitutively in low amounts, and mostly its de-tection has been possible only by indirect means, i.e.,induction of IFN-induced OASE or establishment of anantiviral state (23, 29, 36, 37). When we compared IFNlevels in culture supernatants from C57BL/6, BALB/c, andB6.C-H28c macrophages, we observed significant differ-ences in that BALB/c macrophages consistently secretedbetween 12 and 54 IU of IFN whereas macrophages from thetwo other strains produced low-to-undetectable amounts(Table 1). As expected, the IFN levels increased with time ofculture and thus correlated with the number of maturemacrophages differentiating from myeloid precursor cells inthe presence of M-CSF. The highest titers were found on day7, 1 day after the cultures were supplied with fresh medium.The antiviral activity was completely neutralized by mono-clonal antibodies to IFN-P. No differences were observedbetween crude LCM and purified mouse M-CSF as mediumsupplements (data not shown).We also studied IFN gene expression at the RNA level by

Northern (RNA) blot analysis. As shown in Fig. 1, ahybridization signal was observed for IFN-13 mRNA inBALB/c BMM (lanes C) whereas no expression was detect-able in BMM from C57BL/6 (lanes B) or B6.C-H28c con-genic mice (data not shown). These results were in agree-ment with IFN levels measured in the corresponding culturesupernatants and indicate that expression of endogenous

J. VIROL.

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

IFN REGULATES NDV-INDUCED CYTOKINE GENE EXPRESSION 4841

TABLE 1. Secretion of endogenous IFN inM-CSF-cultured BMM

Rangec of IFN titersCulture conditionsa (IU/ml) with the following

Day of macrophage donor:cultureb

Anti-IFN-a Anti-IFN-P C57BU6 BALB/c B6.C-(NU) (NU) H28C

3 0 0-4 06 0 6-18 0-47 0-4 12-54 0-4

100 3 0 0-4 06 0 6-12 07 2-4 18-60 0-4

100 3 0 0 06 0 0 07 0 0 0

a BMM were cultured for 7 days in RPMI 1640 medium containing 10o FCSand 500 CFU of purified mouse M-CSF per ml with or without antibodies toIFN-a or IFN-3 as indicated.

b Day of removal of culture supernatant.c From five different petri dishes.

IFN in macrophages is under genetic control. Similarly, wealso found endogenous IL-6 gene expression in M-CSF-cultured BALB/c BMM (Fig. 1, lanes C). Interestingly,however, no IFN-P and IL-6 mRNAs were visible in RNApreparations from M-CSF-cultured macrophages kept in thepresence of anti-IFN-B (Fig. 1, lanes C). This finding sug-gests that constitutive secretion of endogenous IFN-,B in

CONTROL

BALB/c macrophages is maintained by a positive feedbackmechanism.

Autocrine biological effects of endogenous IFN. (i) Cellgrowth inhibition. Since IFN has antiproliferative effects onhematopoietic cells, we compared the influence of endoge-nous IFN on M-CSF-driven growth of macrophages fromC57BL/6, BALB/c, and congenic B6.C-H28c mice. Culturesof bone marrow cells were set up in the presence of 500 CFUof mouse M-CSF per ml with or without addition of 100 NUof monoclonal anti-IFN-a or anti-IFN-P. The cell numberper petri dish on day 7 revealed a significantly reducedgrowth rate in BALB/c BMM cultures (Table 2). Addition ofmonoclonal anti-IFN-P, however, caused a pronounced in-crease in BALB/c macrophage counts but had only littleinfluence on cell growth in C57BL/6 and B6.C-H28c BMMcultures. The observed reduced growth rate of BALB/cmacrophages in M-CSF thus does not reflect a lower sensi-tivity to M-CSF but rather is due to enhanced secretion ofendogenous IFN-P.

(ii) Induction of OASE activity. It has been shown thatendogenous IFN is responsible for elevated levels of OASEin cultures of BMM (23, 29, 37). In view of the observeddifferences in endogenous IFN levels, we also determinedthe levels of OASE in BMM from all three of the mousestrains under study. As can be seen additionally from Table2, there was clear-cut induction of the enzyme in M-CSF-cultured control macrophages from all of the mouse strainscompared with those treated with anti-IFN-P. Addition ofmonoclonal anti-IFN-a had no effect. OASE activity wasalso more elevated in macrophages from BALB/c mice thanin those from C57BL/6 and B6.C-H28c mice, probably also

NDV

IFN-a IFN-f IL- 6Mousestrain: CCBB CCBB CCBB

anti-IFNP: + - + - +- +- + - +-

28 s - _ _8 s -. - -

j-ACTIN

IFN-a IFN-f IL-6

CCBB CCBB CCBB

+_-+- ++- -_+ _+

13- ACTIN

2BS _-, _---v 1,1" -9.loil

185s *" ---5~ _~

FIG. 1. Endogenous and NDV-induced cytokine gene expression in BMM from C57BL/6 and BALB/c mice. Bone marrow cells werecultured in mouse M-CSF (500 CFU/ml) with or without addition of 100 NU of anti-IFN-,B. Total RNA was extracted from 7-day-old culturesof pure BMM, either noninfected (control) or 8 h after infection with NDV (100 HAU/106 cells). Agarose gels with 10 ,ug of RNA per lanewere run, blotted onto nylon filters, and hybridized with the indicated cDNA probes. Filters hybridized to IFN-a probes were rehybridizedfirst to IL-6 cDNA and subsequently to 13-actin cDNA. Filters hybridized to the IFN-P probe were rehybridized to ,B-actin cDNA only. Mousestrains: C, BALB/c; B, C57BL/6. The exposure times for the filters were 96 h for controls, 16 h for B NDV, 40 h for C NDV, and 8 h for1-actin.

VOL. 65, 1991

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

4842 ZAWATZKY ET AL.

TABLE 2. Secretion of endogenous IFN by M-CSF-cultured BMM and its effect on cell growth, virus replication, and OASE activity

Mouse Culture conditionsa VSV replication Macrophage growth OASE activitystrain Anti-lFN-a Anti-IFN-p (PFU/culture)b (cell no., 106) oligo(A) mgf h-li]d

C57BL/6 - - 3.40 1.25 20,620+ - 3.8 1.30 12,800- + 6.18 1.35 1,500

BALB/c - - 3.08 0.76 98,000+ - 3.60 0.90 92,000- + 6.40 1.25 980

B6.C-H28C - - 3.60 1.20 NDe+ - 3.71 1.25 ND- + 6.3 1.35 ND

a BMM were cultured for 7 days in RPMI 1640 medium containing 10%o FCS and 500 CFU of purified mouse M-CSF per ml with or without 100 NU ofanti-IFN-a or anti-IFN-P as indicated.

b VSV titers are given as logl0 values.Cell numbers are means of six petri dishes. All assays were carried out on day 7 of culture.

d Enzyme activity was measured in cell extracts.eND, not done.

because of the higher levels of endogenous IFN releasedfrom BALB/c macrophages.

(iii) Induction of resistance to VSV infection. Cultures ofBMM set up in parallel were used to study the antiviral effectof endogenous IFN on infection by VSV. In M-CSF-culturedmacrophages from all three mouse strains, VSV replicationwas low, independently of the observed endogenous IFNlevels in the culture supernatants prior to infection (Table 2).Anti-IFN-a treatment of macrophages caused a slight (five-fold) rise in VSV titers, whereas in anti-IFN-p-treated cellsthere was a 3-order-of-magnitude increase in virus yield.Addition of 20 IU of rIFN-a4 to these cultures 18 h prior toinfection, however, induced a comparable antiviral state inmacrophages from all three mouse strains (data not shown),indicating no differences in the sensitivities of BMM to theantiviral effects of IFN.

(iv) No change in expression of macrophage maturationmarkers. Since IFN possesses multiple immunomodulatoryeffects, we investigated whether endogenous IFN in BMMinfluenced M-CSF-driven differentiation by monitoring theexpression of the surface antigens Mac-1 and F4/80 (4, 40).The Mac-1 glycoprotein complex (CD11b/CD18) is a mac-rophage-granulocyte maturation marker which seems to beinvolved in degranulation and cellular adhesion processes (3,39). F4/80 is a specific marker for mouse mononuclearphagocytes (25). Macrophages from 8-day-old cultures con-taining 500 CFU ofM-CSF per ml with or without addition ofmonoclonal anti-IFN-P were analyzed by fluorescence-acti-vated cell sorter. In both groups, 88 to 96% of the cellsshowed specific staining for either antigen. In addition, morethan 90% of all macrophages stained for the macrophage-specific enzyme a-naphthyl-esterase and more than 95% ofthe cells ingested heat-inactivated Saccharomyces cerevi-siae particles (data not shown).

Specific regulation by endogenous IFN of NDV-inducedcytokine production. Production of IFN can be enhanced incell cultures by pretreatment of the cells with low doses ofIFN, a mechanism termed priming (18, 44). We were there-fore interested to see whether endogenous IFN also exerteda priming effect on subsequent NDV-induced IFN produc-tion in BMM from C57BL/6 and BALB/c mice as high andlow responders, respectively. BMM were obtained by cul-ture in the presence ofM-CSF plus neutralizing antibodies to

IFN-ca or IFN-P and induced on day 7 with NDV or HSV.The results are depicted in Table 3. To our surprise, neutral-ization of endogenous IFN by anti-IFN-P pretreatment didnot suppress but markedly enhanced subsequent IFN pro-duction in response to NDV, whereas anti-IFN-oa had noeffect. The enhancing effect was much more pronounced inmacrophages from BALB/c mice, resulting in nearly com-plete abrogation of the genetically determined differences inIFN production between the two mouse strains. The effectwas also specific for NDV, since pretreatment of macro-phages with anti-IFN-, had little influence on HSV-inducedIFN levels.

In addition, production of IL-6 in response to NDVseemed to be controlled by similar host genes, sinceC57BL/6 BMM again displayed the high-producer pheno-type and endogenous IFN also differentially influenced IL-6levels in BMM from high- and low-producer mouse strains(Table 3).

TABLE 3. Virus-induced production of IFN and IL-6 inM-CSF-cultured BMM: regulation of virus-induced

cytokine release by endogenous IFN

Preculture IFN concn (IU/mI) IL-6 concn (U/ml)condition'

Inducer

IFN-t IFN-t C57BL/6 BALB/c C57BL/6 BALB/c

- - NDV 2,400 240 1,800 60+ - NDV 2,700 300 1,500 60- + NDV 16,200 8,400 3,600 2,500+ + NDV 16,200 9,600 3,200 1,800

- - HSV 280 75 NDb ND+ - HSV 210 90 ND ND- + HSV 320 180 ND ND+ + HSV 400 210 ND ND

a Macrophages were cultured for 7 days in purified mouse M-CSF alone(500 CFU/ml) or M-CSF with or without 100 NU of either monoclonalanti-IFN-a or monoclonal anti-IFN-j. The cells were then transferred to24-well plates at 2.5 x 105 per well and induced with NDV (125 HAU/2.5 x105 cells) or HSV at a multiplicity of 2. Culture supernatants were harvestedfor cytokine assays 18 h later.

b ND, not done.

J. VIROL.

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

IFN REGULATES NDV-INDUCED CYTOKINE GENE EXPRESSION 4843

Similar results were obtained when the amounts of NDV-induced IFN-a, IFN-,, and IL-6 gene expression were

analyzed on the RNA level by Northern blot hybridization(Fig. 1). This is in agreement with earlier results showingthat NDV-induced IFN production is regulated at the levelof transcription (24). RNA was extracted from macrophages8 h after NDV infection, when maximal levels of IFN-a,IFN-P, and IL-6 mRNAs had accumulated in macrophagesfrom all of the mouse strains tested, as analyzed in controlexperiments (data not shown) and in previous studies (23,24). As expected, steady-state levels of NDV-induced IFNand IL-6 mRNAs were much higher in C57BL/6 BMM thanin BALB/c BMM when precultured in M-CSF alone. It isalso obvious that these cells almost exclusively expressedIFN-, mRNA, whereas IFN-a mRNA was hardly detect-able. Preculture in M-CSF plus anti-IFN-,B, however, notonly caused a strong increase in IFN-P mRNA levels inresponse to NDV but, in particular, activated IFN-a gene

expression in BMM from both C57BL/6 and BALB/c mice.Characterization of the corresponding culture supernatantsrevealed a good correlation with mRNA levels: M-CSF-cultured macrophages produced mainly IFN-P in responseto NDV, whereas a composition of about 50% IFN-a and50% IFN-p was found in NDV-induced BMM precultured inM-CSF plus anti-IFN-P (see Table 5). Analysis of NDV-induced IL-6 mRNA levels in C57BL/6 and BALB/c macro-

phages revealed strong parallels to the regulation of IFN-Pgene expression. Again, there was a dramatic increase insteady-state levels of IL-6 mRNA in BALB/c macrophagesprecultured in the presence of anti-IFN-P, whereas inC57BL/6 macrophages, IL-6 mRNA levels remained almostunaffected.Enhanced sensitivity to a particular biological effect of IFN

encoded by the If-i' locus. Since we had observed higherlevels of endogenous IFN in BMM from BALB/c mice, wewanted to know whether the NDV-specific low-responderphenotype of these cells resulted from dose-dependentdownregulation of cytokine gene expression by endogenousIFN. To this end, we made use of the congenic line B6.C-H28c. BMM from these mice produced very low or unde-tectable amounts of endogenous IFN comparable to those ofBMM from the C57BL/6 parental strain (Table 1). Levels ofNDV-induced IFN, in contrast, were close to those ofBALB/c BMM (Table 4), since B6.C-H28c macrophagesexpress the If-II allele (15). However, preculture of B6.C-H28C macrophages in M-CSF plus anti-IFN-1 resulted inenhancement of the IFN response to NDV infection similarto that observed for BALB/c. This indicates that the differ-ences in endogenous IFN production are not responsible forthe low-responder phenotype of these cells. The data inTable 4 also reveal that overnight exposure to 20 IU ofrIFN-a4 in the presence of anti-IFN-P also mediated stronginhibition of NDV-induced IFN production in B6.C-H28cand BALB/c macrophages, which both carry the If-i' allele,whereas this inhibition was only marginal in C57BL/6 mac-

rophages expressing the If-lh allele. The inhibitory effectmediated by exogenous or endogenous IFN was also specificfor NDV, since pretreatment of BMM with anti-IFN-,caused little changes in the HSV-induced IFN response, andaddition of 20 IU rIFN-a4 prior to HSV infection evenresulted in slight enhancement of IFN titers (Table 4).

In addition, Northern blot analysis was performed inparallel with RNAs from C57BL/6 and congenic B6.C-H28cmacrophages 8 h after NDV infection, and the results are

depicted in Fig. 2. Steady-state levels of IFN-a, IFN-1, andIL-6 mRNAs were considerably enhanced by pretreatment

TABLE 4. Regulation by endogenous and exogenous IFN ofNDV- and HSV-induced IFN production in mouse macrophages

Preculture conditions' IFN titer (IU/ml) at 18 h postinduction with thefollowing macrophage donor:

Anti- Anti- Amt of C57BL/6 BALB/c B6.C-H28cAFNoti AFNt- rIFN-a____4__IFN-cs IFN-,B (IU) NDV HSV NDV HSV NDV HSV

- - 2,400 350 180 40 250 320+ - 2,100 450 180 72 320 320- + 16,200 320 9,600 96 10,800 360- + 20 5,400 840 30 110 250 480- + 200 4,800 260 <3 24 90 220

a Macrophages were cultured for 7 days in mouse M-CSF (500 CFU/ml)with or without monoclonal anti-IFN-a or anti-IFN-,B (100 NU/ml). Macro-phages were then transferred to 24-well plates at 2.5 x 105 per well, and 20 or200 IU of rIFN-a4 was added to some of the cultures for 20 h in the presenceof anti-IFN-0. The cells were then induced by NDV (125 HAU/2.5 x 105 cells)or HSV at a multiplicity of 2.

of the macrophages with anti-IFN-P (lanes 3) compared withcells cultured in M-CSF alone (lanes 2). The enhancing effectwas much more pronounced for IFN-a than for IFN-, orIL-6 mRNA. Upon exposure of macrophages to 20 IU ofrIFN-a4, we also observed strong inhibition of NDV-in-duced cytokine expression on the RNA level (lanes 4), andthe inhibitory effect was most evident for IFN-a mRNA,which became hardly detectable in these cells. In contrast,there was little change in IFN-,B or IL-6 mRNA expression inNDV-induced BMM from C57BL/6. The culture superna-tants corresponding to these RNA preparations were ana-lyzed in a neutralization assay using monoclonal anti-IFN-aand anti-IFN-,. The results, summarized in Table 5, re-vealed that the relative proportions of IFN-a and IFN-Pwere in agreement with the observed steady-state mRNAlevels.

Influence of anti-IFN-I8 treatment of BMM on NDV repli-cation. Partial primary transcription of the viral genome andformation of double-stranded RNA intermediates after infec-tion with single-stranded RNA viruses are considered onemechanism that leads to efficient IFN induction by viruses(16, 32, 33). We therefore investigated the possibility thatenhanced viral replication accounted for the high NDV-specific IFN response in anti-IFN-,-treated macrophages.No hemagglutination activity, however, was found in culturesupernatants from infected BMM, and similarly, no NDVtranscripts coding for the viral F protein were detected intotal RNA preparations from these cells (data not shown).This indicates that nonpermissiveness of mouse macro-phages to NDV infection is not due to an antiviral stateinduced by endogenous IFN-13.

DISCUSSION

The present report deals with two major independentfindings: (i) macrophages from inbred mouse strains differ inthe capacity to produce endogenous IFN-1 during culture inM-CSF, and (ii) the regulatory effect of the If-] gene locus onNDV-induced cytokine release in macrophages is influencedby endogenous IFN.

Secretion of low amounts of antiviral activity in cultures ofmouse BMM was initially described by Moore et al. (34), andsubsequently, indirect evidence was also obtained for thesecretion of endogenous IFN by macrophages in vivo (6, 20,36). Our studies have documented significant differences inendogenous IFN production between BMM from BALB/c

VOL. 65, 1991

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

4844 ZAWATZKY ET AL.

IFN-ct

Exp. time B.C B1 2 3 4 5 6 7 8

B.C B

1 2 3 4 5 6 7 8

IL-6

BC B

1 2 3 4 5 6 1 8

28l --1 8 _

5hS w *9 to9

28, _

66h 18s- .50 g

5h18!S-99O

FIG. 2. Regulation by the If-] locus of NDV-induced cytokine gene expression in BMM. Total RNA was extracted from BMM andanalyzed by Northern blot hybridization as described in the legend to Fig. 1. Lanes: 1 and 5, noninfected BMM cultured in M-CSF alone;2 and 6, NDV-infected BMM precultured in M-CSF alone; 3 and 7, NDV-infected BMM precultured in M-CSF plus anti-IFN-,; 4 and 8,NDV-infected BMM precultured in M-CSF, anti-IFN-P, and 20 IU of rIFN-a4 for the last 20 h prior to NDV infection. Mouse strains: B,C57BL/6; B.C, B6.C-H28c. Two exposure (Exp.) times of the same filters are presented to give a better impression of the differences in thesteady-state levels of individual cytokine mRNAs.

and C57BL/6 mice. The IFN was identified as IFN-,B, andthe differences between the two mouse strains were visibleon the RNA and protein levels, suggesting that host genesregulate the amount of endogenous IFN gene expression.Among the autocrine effects of endogenous IFN, we ob-served pronounced growth inhibition in cultures of BALB/cBMM compared with C57BL/6 BMM, and this was probablydue to the higher levels of endogenous IFN, since in thepresence of monoclonal anti-IFN-,, proliferation rates wereidentical in the two cultures. It has also been reported thatBMM from BALB/c mice exhibit higher sensitivity to theantiproliferative effect of exogenously added IFN-a/I (11),and this higher sensitivity may further enhance the observedgrowth-inhibitory effect in these cells. On the other hand, nodifferences in IFN-mediated antiviral resistance to VSV wasdetectable among BMM from C57BL/6, B6.C-H28c, andBALB/c mice. This observation may be explained by thefact that growth inhibition in general requires larger amountsof IFN than does establishment of an antiviral state (45).When we studied the regulation of cytokine release in

TABLE 5. Variation of NDV-induced levels of mouse IFN-a andIFN-P in macrophages

11% % IFNCells IFN-a IFN-P titera

Peritoneal macrophages <10 >90 8-15

BMM cultured in:M-CSF <10 >90 8-15M-CSF + anti-IFN-P 40-50 50-60 1.8-2.5M-CSF + anti-IFN-0 + rIFN-ot <10 >90 8-20a If-lh/If-il (high- versus low-responder) ratio.

response to NDV, we found that BMM from BALB/c andB6.C-H28C mice secreted considerably lower levels of IFNand IL-6 than did BMM from C57BL/6 mice that were highproducers of both cytokines. Although coexpression ofIFN-, and IL-6-for which the term IFN-P2 had been usedpreviously-following induction by virus is a known phe-nomenon (7, 30), these data additionally indicate that the If-llocus regulates the production of both cytokines in responseto NDV. In addition, we made the striking observation thatendogenous IFN-P caused not priming but blocking ofIFN-a, IFN-P, and IL-6 gene expression after NDV infec-tion of macrophages. The blocking effect was much morepronounced in BMM from BALB/c and B6.C-H28c micethan in those from C57BL/6 mice. This indicates that theamount of blocking does not correlate with the level ofendogenous IFN secreted by these macrophages but ratherwith their If-i' genotype, since BMM from congenic B6.C-H28C mice also produced low or undetectable levels ofendogenous IFN-P, like cells from the C57BL/6 parentalstrain. This hypothesis was further supported by the findingthat addition of identical amounts of mouse rIFN-a4 in thepresence of monoclonal anti-IFN-,B also exerted a suppres-sive effect on subsequent NDV-induced cytokine produc-tion, which was much more pronounced in BALB/c andB6.C-H28c BMM with the If-i' genotype.Taken together, there is evidence that the If-i locus

contains a gene(s) which is activated by IFN and that the twoalleles respond differently to endogenous IFN-P or exoge-nous IFN-a. Considering specific inhibition of cytokine geneexpression as the only function attributable to these genes upto now, we propose that genes within the If-i' allele displayhigher activity when stimulated by IFN.We can only speculate about the mechanism by which

IFN-activated genes within the If-i locus regulate cytokine

"I _

3-ACTIN

_ o

J. VIROL.

09

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

IFN REGULATES NDV-INDUCED CYTOKINE GENE EXPRESSION 4845

gene expression in macrophages. In view of the specificityfor NDV, the possibility that If-i-derived gene products ingeneral downregulate IFN and IL-6 production can beexcluded. They seem, rather, to interfere with NDV-specificelements responsible for cytokine induction. In our study,replication of the viral genome as assessed by F proteincDNA could not be monitored in macrophages deprived ofendogenous IFN-P. It may be, however, that under theseconditions primary transcription of viral genes occurs, lead-ing to formation of double-stranded RNA intermediateswhich, in turn, induce IFN production. Activation by endog-enous IFN of either If-i' or If-Ph would consequently resultin more- or less-pronounced inhibition of viral transcriptionand double-stranded RNA formation, depending on the If-]allele involved.The physiological role of If-] remains to be elucidated.

Since the natural hosts of NDV are fowl, and mice are notsusceptible to this virus, it seems likely that the assumedinterference of mouse If-i gene products with NDV whichleads to downregulation of cytokine induction in host cellsdoes not represent its only function. Highly significantindividual variations in NDV-induced IFN titers have alsobeen observed in cultures of human peripheral blood leuko-cytes (5), suggesting that regulatory gene loci similar to If-iare also present in the human genome. This finding points toa possible conservation of If-] during evolution and alsoargues for additional biological functions of its gene prod-ucts.

Concerning virus-specific action and inducibility by IFN,some interesting parallels between If-i and the mouse Mxgene emerge. This gene confers resistance to influenza virusinfection and is inducible by IFN-ot/, (for a review, seereference 41). Its functional allele, Mx+, encodes a proteinwhich accumulates in the nucleus and specifically interfereswith influenza virus replication (28). Meanwhile, homologsof the mouse Mx protein have also been identified in humancells and their inducibility by IFN has been proven (2, 19). Inthis respect, further studies concerning the alleles of If-i willalso be of significant interest for the characterization of oneor more new genes activated differentially by endogenousIFN.

ACKNOWLEDGMENTSWe are indebted to 0. Haller for helpful discussions, Y. Watanabe

for providing hybridoma lines 4EA1 and 7FD3, and E. R. Stanley forthe gift of purified mouse M-CSF. The DNA probes for mouseIFN-a1, IFN-a2, IFN-a4, IFN-p, and IL-6 were kindly provided byE. C. Zwarthoff, Y. Kawade, and J. van Snick. We also thank P. T.Emmerson for the NDV F-protein cDNA. The excellent secretarialassistance of M. Kulka is gratefully acknowledged.

This work was supported in part by a grant from the DeutscheForschungsgemeinschaft (Za 98/2-2).

REFERENCES1. Aarden, L. A., E. R. De Groot, 0. L. Schaap, and P. M.

Landsdorp. 1987. Production of hybridoma growth factor byhuman monocytes. Eur. J. Immunol. 17:1411-1416.

2. Aebi, M., J. Fah, N. Hurt, C. E. Samuel, D. Thomis, L.Bazzigher, J. Pavlovic, 0. Haller, and P. Staeheli. 1989. cDNAstructures and regulation of two interferon-induced human Mxproteins. Mol. Cell. Biol. 9:5062-5072.

3. Anderson, D. C., L. J. Miller, F. C. Schmalstieg, R. Rothlein,and T. A. Springer. 1986. Contributions of the Mac-1 glycopro-tein family to adherence-dependent granulocyte functions:structure-function assessments employing subunit-specificmonoclonal antibodies. J. Immunol. 137:15-27.

4. Austyn, J. M., and S. Gordon. 1981. F4/80, a monoclonalantibody directed specifically against the mouse macrophage.

Eur. J. Immunol. 11:805-815.5. Bacon, T. H., A. De Vere-Tyndall, D. A. J. Tyrell, A. M.

Denman, and B. M. Ansell. 1983. Interferon system in patientswith systemic juvenile chronic arthritis: in vivo and in vitrostudies. Clin. Exp. Immunol. 54:23-30.

6. Belardelli, F., F. Vignaux, E. Proietti, and I. Gresser. 1984.Injection of mice with antibody to interferon renders peritonealmacrophages permissive for vesicular stomatitis virus and en-cephalomyocarditis virus. Proc. Natl. Acad. Sci. USA 81:602-606.

7. Billiau, A., J. Van Damme, J. Ceuppens, and M. Baroja. 1989.Interleukin-6, a ubiquitous cytokine with paracrine as well asendocrine functions. Colloq. INSERM 179:133-142.

8. Boursnell, M. E. G., P. F. Green, J. I. A. Campbell, A. Deuter,R. W. Peters, F. M. Tomley, A. C. R. Samson, P. Chambers,P. T. Emmerson, and M. M. Binns. 1990. Insertion of the fusiongene from Newcastle disease virus into a nonessential region inthe terminal repeats of fowlpox virus and demonstration ofprotective immunity induced by the recombinant. J. Gen. Virol.71:621-628.

9. Chambers, P., N. S. Millar, R. W. Bingham, and P. T. Emmer-son. 1986. Molecular cloning of complementary DNA to New-castle disease virus, and nucleotide sequence analysis of thejunction between the genes encoding the hemagglutinin-neur-aminidase and the large protein. J. Gen. Virol. 67:475-486.

10. Chirgwin, J. M., A. E. Przybyla, R. J. MacDonald, and W. J.Rutter. 1979. Isolation of biologically active ribonucleic acidfrom sources enriched in ribonuclease. Biochemistry 18:5294-5299.

11. Dandoy, F., E. De Maeyer, and J. De Maeyer-Guignard. 1981.Antiproliferative action of interferon on murine bone-marrowderived macrophages is influenced by the genotype of themarrow donor. J. Interferon Res. 1:263-270.

12. De Maeyer, E., and J. De Maeyer-Guignard. 1979. Consider-ations on mouse genes influencing interferon production andaction, p. 75-100. In I. Gresser (ed.), Interferon I. AcademicPress, Inc. (London), Ltd., London.

13. De Maeyer, E., and J. De Maeyer-Guignard. 1988. Interferonsand other regulatory cytokines. John Wiley & Sons, Inc., NewYork.

14. De Maeyer, E., J. De Maeyer-Guignard, and D. W. Bailey. 1975.Effect of mouse genotype on interferon production. Linescongenic at the If-1 locus. Immunogenetics 1:438-445.

15. De Maeyer, E., M. C. Hoyez, J. De Maeyer-Guignard, and D. W.Bailey. 1979. Effect of mouse genotype on interferon produc-tion. III. Expression of If-1 by peritoneal macrophages in vitro.Immunogenetics 8:257-263.

16. Falcoff, R., and E. Falcoff. 1970. Induction de la synthesed'interfdron par des RNA bicatenaires. II. Etude de la formeintermediaire de replication du virus Mengo. Biochim. Biophys.Acta 199:147-158.

17. Feinberg, A. P., and B. Vogelstein. 1983. A technique forr1adiolabeling DNA restriction endonuclease fragments to highspecific activity. Anal. Biochem. 132:6-13.

18. Fujita, T., and S. Ohno. 1981. Studies on interferon priming:cellular response to viral and nonviral inducers and requirementof protein synthesis. Virology 112:62-69.

19. Goetschy, J.-F., H. Zeller, J. Content, and M. A. Horisberger.1989. Regulation of the interferon-inducible IFI-78K gene, thehuman equivalent of the murine Mx gene, by interferons,double-stranded RNA, certain cytokines, and viruses. J. Virol.63:2616-2622.

20. Gresser, I., F. Vignaux, F. Belardelli, M. G. Tovey, and M.-T.Maunoury. 1985. Injection of mice with antibody to mouseinterferon /al decreases the level of 2'-5' oligoadenylate syn-thetase in peritoneal macrophages. J. Virol. 53:221-227.

21. Higashi, Y., Y. Sokowa, Y. Watanabe, Y. Kawade, S. Ohno, C.Takaoka, and T. Taniguchi. 1983. Structure and expression of a

cloned cDNA for mouse interferon-beta. J. Biol. Chem. 258:9522-9529.

22. Hirsch, S., J. M. Austyn, and S. Gordon. 1981. Expression of themacrophage-specific antigen F4/80 during differentiation ofmouse bone marrow cells in culture. J. Exp. Med. 154:713-725.

VOL. 65, 1991

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from

4846 ZAWATZKY ET AL.

23. Hoss, A., E. C. Zwarthoff, and R. Zawatzky. 1989. Differentialexpression of interferon alpha and beta induced with Newcastledisease virus in mouse macrophage cultures. J. Gen. Virol.70:575-589.

24. Hoss-Homfeld, A., E. C. Zwarthoff, and R. Zawatzky. 1989. Celltype specific expression and regulation of murine interferonalpha and beta genes. Virology 173:539-550.

25. Hume, D. A., and S. Gordon. 1983. Mononuclear phagocytesystem of the mouse defined by immunohistochemical localiza-tion of antigen F4/80. J. Exp. Med. 157:1704-1709.

26. Kawade, Y., and Y. Watanabe. 1987. Characterization of ratmonoclonal antibodies to mouse interferon alpha and beta, p.197-201. In K. Cantell and H. Schellekens (ed.), Proceedings ofthe Third International TNO Meeting on the Biology of theInterferon System. The Biology of the Interferon System.Nijhoff, The Hague, The Netherlands.

27. Knight, M., P. J. Cayley, R. H. Silverman, D. H. Wreschner,C. S. Guilbert, R. E. Brown, and I. M. Kerr. 1980. Radioim-mune, radiobinding and HPLC analysis of 2-5 A and relatedoligonucleotides from intact cells. Nature (London) 288:189-192.

28. Krug, R. M., M. Shaw, B. Broni, G. Shapiro, and 0. Haller.1985. Inhibition of influenza viral mRNA synthesis in cellsexpressing the interferon-induced Mx gene product. J. Virol.56:201-206.

29. Kruse, A., H. Kirchner, R. Zawatzky, and I. Domke-Opitz. 1989.In vitro development of bone marrow derived macrophages.Influence of mouse genotype on response to CSF and autocrineIFN induction. Scand. J. Immunol. 30:731-740.

30. Le, J., and J. Vilcek. 1989. Interleukin 6: a multifunctionalcytokine regulating immune reactions and the acute phaseprotein response. Lab. Invest. 61:588-602.

31. Lehrach, H., D. Diamond, J. M. Wozney, and H. Boedtker. 1977.RNA molecular weight determinations by gel electrophoresisunder denaturing conditions: a critical reexamination. Biochem-istry 16:4743-4751.

32. Marcus, P., and F. J. Fuller. 1979. Interferon induction byviruses. II. Sindbis virus interferon induction requires one-quarter of the genome-genes G and A. J. Gen. Virol. 44:169-177.

33. Marcus, P. I. 1984. Interferon induction by viruses: double-stranded ribonucleic acid as the common proximal inducermolecule. Interferon 3:113-175.

34. Moore, R. N., H. S. Larsen, D. W. Horohov, and B. T. Rouse.1984. Endogenous regulation of macrophage proliferative ex-pansion by colony-stimulating factor-induced interferon. Sci-ence 223:178-180.

35. Mosmann, T. 1983. Rapid colorimetric assay for cellular growthand survival: application to proliferation and cytotoxicity as-says. J. Immunol. Methods 65:55-63.

36. Proietti, E., S. Gessani, F. Belardelli, and I. Gresser. 1986.Mouse peritoneal cells confer an antiviral state on mouse cell

monolayers: role of interferon. J. Virol. 57:456-463.37. Resnitzky, D., A. Yarden, D. Zipori, and A. Kimchi. 1986.

Autocrine beta-related interferon controls c-myc suppressionand growth arrest during hematopoietic cell differentiation. Cell46:31-40.

38. Russel, P. H. 1984. Newcastle disease virus: the effect ofmonoclonal antibody in the overlay on virus penetration and theimmunoselection of variants. J. Gen. Virol. 65:795-798.

39. Schleifenbaum, B., R. Moser, M. Patarroyo, and J. Fehr. 1989.The cell surface glycoprotein Mac-1 (CD11b/CD18) mediatesneutrophil adhesion and modulates degranulation independentlyof its quantitative cell surface expression. J. Immunol. 142:3537-3545.

40. Springer, T., G. Galfre, D. S. Secher, and C. Milstein. 1979.Mac-1: a macrophage differentiation antigen identified by mono-clonal antibody. Eur. J. Immunol. 9:301-306.

41. Staeheli, P., and 0. Haller. 1987. Interferon-induced Mx pro-tein: a mediator of cellular resistance to influenza virus. Inter-feron 8:1-23.

42. Stanley, E. R., and L. J. Guilbert. 1981. Methods for thepurification, assay, characterization and target cell binding of acolony stimulating factor (CSF-1). J. Immunol. Methods 42:253-284.

43. Stark, G. R., R. E. Brown, and I. M. Kerr. 1981. Assay of(2'-5')-oligoadenylic acid synthetase levels in cells and tissues: aconvenient poly (I)-poly (C) paper-bound enzyme assay. Meth-ods Enzymol. 79:194-199.

44. Stewart, W. E., II, L. B. Gosser, and R. Z. Lockart, Jr. 1971.Priming: a nonantiviral function of interferon. J. Virol. 7:792-801.

45. Taylor-Papadimitriou, J., F. Balkwill, N. Ebsworth, and E.Rozengurt. 1985. Antiviral and antiproliferative effects of inter-feron in quiescent fibroblasts are dissociable. Virology 147:405-412.

46. Van Heuvel, M., I. J. Bosveld, A. T. A. Mooren, J. Trapman,and E. C. Zwarthoff. 1986. Properties of natural and hybridmurine alpha interferons. J. Gen. Virol. 67:2215-2222.

47. Van Snick, J., S. Cayphas, J.-P. Szikora, J.-L. Renauld, E. VanRoost, T. Boon, and R. J. Simpson. 1988. cDNA cloning ofmurine interleukin-HP1: homology with human interleukin-6.Eur. J. Immunol. 18:193-200.

48. Zawatzky, R., I. Gresser, E. De Maeyer, and H. Kirchner. 1982.The role of interferon in the resistance of C57BL/6 mice tovarious doses of herpes simplex virus type 1. J. Infect. Dis.146:405-410.

49. Zawatzky, R., H. Kirchner, J. De Maeyer-Guignard, and E. DeMaeyer. 1982. An X-linked locus influences the amount ofcirculating interferon induced in the mouse by herpes simplexvirus type 1. J. Gen. Virol. 63:325-332.

50. Zwarthoff, E. C., A. T. A. Mooren, and J. Trapman. 1985.Organization, structure and expression of murine interferonalpha genes. Nucleic Acids Res. 13:791-804.

J. VIROL.

at RE

GE

NS

BU

RG

on August 4, 2009

jvi.asm.org

Dow

nloaded from